A digital twin of the coupled climate, economic and financial systems

Guest blog by Dr Lara Stoimenova and Dr Craig Mounfield, Founders of Simulnova Software Limited

The challenge: understanding the potential impact of climate change on our economy and our financial system is very difficult using traditional methods

Climate change is happening all around us: warming oceans, melting polar glaciers, raising sea levels, more extreme weather events are but some of the examples.

Climate change does not only impact our planet, but also our economy and financial system. Both physical risk (for example, acute storm damage) and transition risk (for example, climate related taxes) can lead to shocks in the UK economy. For example, firms’ assets can be impaired or stranded leading to potential losses to the value of their business. This in turn can impact on financial instruments (such as equity options) that reference these firms and are traded by financial institutions. Thus climate shocks can have a direct impact on the stability of financial markets.

However the climate, economic and financial systems are highly complex. They are composed of large numbers of discrete, heterogeneous agents (consumers, firms, banks, governments etc.) that interact with one another at a microscopic level. The multitude of these interactions can create feedback loops leading to emergent, aggregate macroeconomic and financial behaviour which is often unpredictable (such as business cycles, recessions, and financial crashes).

Because of the inherent difficulties in characterising the interactions between the climate, economic and financial systems a complete, predictive model of the coupled systems is very difficult to construct.

The solution: building a digital twin of the coupled economic, financial and climate system can help understand and mitigate future risks

Using Agent Based Computational Economics (ACE),we have constructed a digital twin of the coupled economic, financial and climate systems (the system). ACE assumes we can view the economy as a complex system where the economy and financial systems can be modelled as a dynamic composition of interacting agents where we can:

model individual (not aggregated) heterogeneous agents (firms, consumers, banks etc.);
assign bespoke behavioural heuristics to each agent; and
capture the detailed interaction mechanisms between the agents and hence the strong dynamic feedback loops that can form between them.

From the multitude of microscopic interactions between the agents we observe emergent macroscopic behaviour (such as business cycles, recessions and financial crashes). The ensuing endogenous emergent behaviour has the potential to be very rich and, in many cases, a priori unanticipated (including the existence of discontinuous tipping points, which in a climate context can be very important to understand).

Our approach can be used to understand, for example:

What is the impact of exogenous shocks (at the level of individual agents) on the dynamic trajectory of the system?
What are the detailed transmission pathways by which shocks can propagate through the system?
Counterfactual analysis e.g. “what would happen if …?”
What is the likely impact of new rules and regulations e.g. carbon taxes, bank capital buffers?

Case study: an economy hit by a severe heat wave

Our digital twin is composed of a variety of agents as shown in the table below:

The agents are endowed with simple behavioural rules that determine how they interact with and respond to each other. For example, the Central Bank varies the base rate to hit its target inflation rate using a simple Taylor rule.

One of the scenarios we have modelled involves the economy being hit by an acute physical shock in the form of a severe heatwave that persists for one month. We assume the impact on the economy to be:

A sharp drop in consumer demand; for example, consumers staying indoors and thus not spending cash.
A drop of firms productive output; for example, firms having to close factories or scale back production due to the excessive heat.

As the shock dynamically propagates through the economy, we observe unemployment rising, matched by an increase in firms’ default rates etc., before reverting back to near to its long-term trend behaviour.

The shock, even though short lived (one month), induced a strong response from firms in terms of cutting their workforces to adapt to the new (reduced) levels of demand from consumers. This would most likely increase political pressure on the government to act. Thus, we have then assumed that the government intervenes by introducing a temporary tax (12 months after the initial shock, lasting for 12 months) to support consumers during a period of high unemployment: this tax is collected from firms based on their assumed contribution to total GHG emissions (the polluter pays principle).

The outputs of our analysis are shown in the Figure below. In particular, we observe that the impact of the fiscal response is to induce a second wave of firm defaults (because of the increased tax burden) resulting in a second spike in unemployment. Furthermore, even when the temporary tax is removed the level of unemployment can remain at a higher level (compared to pre-shock ) for an extended period of time.

We note that our case study outlines a scenario where we have applied a very extreme shock for illustrative purposes to show the capability of our approach.

A more detailed description of our approach can be found on the SSRN platform.

Conclusion

Our case study shows that building a digital twin of the coupled economic, financial and climate systems can be a very valuable tool for firms and regulators to better understand and mitigate the impacts of climate related shocks to these systems.

If you would like to discuss our approach please send a message to Lara Stoimenova over LinkedIn

Digital Twins updates

Sign-up to receive updates on techUK’s work around supporting the adoption of digital twin technology throughout our economy. You’ll receive regular insights, policy updates, event invitations and projects you can get involved with.